A new method was developed to estimate complex multivariate volatility models more efficiently. Instead of estimating all variables together, the method breaks it down into simpler steps. First, it estimates the volatility of individual returns, then it estimates the correlations between them. This approach is easier and faster than traditional methods. The accuracy of the first-step estimator was proven to be reliable. The method was also successful in testing specific restrictions in the model. An example using financial data showed the effectiveness of this new approach.